algorithmic-trading-anthology-july-2026-regime-transition-rather-than-regime-level

Conclusion: Null

No effect at any horizon tested, across 18 tests.

Every test, against the significance threshold

persistence-robust p · left is stronger evidence

Filled markers fall below the 0.05 threshold before multiple-testing correction; hollow markers do not. Correction is applied across the whole batch, so an individual marker clearing the line is not a result on its own.

Tests

ticker × horizon

p < 0.05

uncorrected

Best q

1.00

BH-FDR

Survived

q ≤ 0.10

Method

Hypothesis

A standing El Niño is public information and should already sit in the curve. The tradeable event, if one exists, is the change in state.

Rule

signal = soi_chg3 # 3-month change

prior = -1 # lower SOI -> higher returns

function test(instrument, horizon h):

y = sum(returns[instrument], t+1 .. t+h) # strictly forward

x = signal[t] # known at t

beta, t_stat = OLS(y ~ x) with Newey-West HAC, lag = h + 1

# null: slide x against y by every admissible offset, so

# persistence appears identically under the null

null = [slope(y, roll(x, k)) for k in all offsets]

p = (count(abs(null) >= abs(beta)) + 1) / (len(null) + 1)

return beta, t_stat, p

reject if Benjamini-Hochberg q <= 0.10 across the whole study

Evaluation

Table 1 · Every test in this hypothesis

forward return regressed on the SOI signal

Instrument	Horizon	n	β	HAC t	Shift p	Sign as predicted
ZW=F	1m	310	-0.0001	-0.12	0.465	yes
ZW=F	3m	308	+0.0015	+0.77	0.815	no
ZW=F	6m	305	+0.0037	+1.69	0.933	no
SB=F	1m	314	+0.0008	+0.90	0.839	no
SB=F	3m	312	+0.0026	+1.10	0.862	no
SB=F	6m	309	+0.0044	+1.37	0.909	no
KC=F	1m	316	+0.0005	+0.63	0.738	no
KC=F	3m	314	+0.0022	+1.37	0.866	no
KC=F	6m	311	+0.0041	+1.50	0.889	no
CC=F	1m	316	+0.0001	+0.19	0.595	no
CC=F	3m	314	+0.0014	+0.69	0.747	no
CC=F	6m	311	-0.0007	-0.24	0.391	yes
ZM=F	1m	312	+0.0003	+0.38	0.645	no
ZM=F	3m	310	+0.0023	+1.19	0.847	no
ZM=F	6m	307	+0.0068	+2.86	1.000	no
DBA	1m	232	+0.0004	+0.92	0.824	no
DBA	3m	230	+0.0028	+2.09	0.995	no
DBA	6m	227	+0.0056	+2.63	1.000	no

+ 14 more rows− fewer rows

Shift p is the persistence-robust p-value and is the one that decides. HAC t is shown to demonstrate how far conventional inference overstates the evidence.

Reference

Instruments

last price and one-month change, live from Yahoo Finance

Symbol	Notes
CC=F	Cocoa — continuous futures on ICE Futures, in USD.
DBA	Invesco DB Agriculture Fund — exchange-traded fund listed on NYSEArca, in USD.
KC=F	Coffee C — continuous futures on ICE Futures, quoted in US cents.
SB=F	Sugar No. 11 — continuous futures on ICE Futures, quoted in US cents.
ZM=F	Soybean Meal — continuous futures on CBOT, in USD.
ZW=F	Chicago SRW Wheat — continuous futures on CBOT, quoted in US cents.

Sample

1970–2026 monthly

Sources

Queensland Department of Agriculture and Fisheries. Southern Oscillation Index. The Long Paddock. Monthly, 1876–. www.longpaddock.qld.gov.au/soi/soi-data-files

Yahoo Finance. Historical market prices. Adjusted close, month-end. Retrieved with yfinance. finance.yahoo.com

Revision history

Rev	Stage	Status	Change
01	Draft	—	Direction and criteria fixed before the experiment ran
02	Tested	Null	18 tests; 0 below p=0.05, none surviving FDR